Physical conditions of the interstellar medium in star-forming galaxies at z~1.5 Abstract Masao Hayashi (NAOJ, Mitaka) Subaru seminar @ Subaru Telescope, NAOJ 10 March 2015 To be re-submitted soon to PASJ C. Ly (NASA), K. Shimasaku, K. Motohara, R. Goto, Y. Naito (Univ. of Tokyo), M. Malkan (ULCA), T. Nagao (Ehime Univ.), N. Kashikawa (NAOJ) Subaru/FMOS near-ir spectroscopy for 118 [OII] emission line galaxies at z~1.5 in the Subaru Deep Field Six nebular emission lines in rest-optical: [OII], Hb, [OIII], Ha, [NII], and [SII] Use the line ratios to investigate physical conditions of the ISM
A period called Cosmic High Noon Stellar mass density: ~10 times from ~5% to ~50% (e.g., Muzzin+2013) Star formation rate density: a peak at z=1-3 (e.g., Hopkins & Beacom 2006) AGN activity: a peak at z=2-3 (e.g., Hopkins+2007) Morphology: from irregular to one classifiable as a Hubble sequence (e.g., Conselice+2005) Muzzin+13 The redshift of z=1-3 is an era of rapid evolution of galaxy properties
Enhanced star formation activity at z>1.5 Tasca+14 z~1.5 TIR LIRG ULIRG Specific SFR (SFR/M*) of galaxy is increasing with redshift higher Goto+11 The contribution of LIRGs and ULIRGs to cosmic SFR density is increasing as redshift is higher Larger contribution of young massive stars to radiation field can alter the condition of the ISM in high-z galaxies, compared to local galaxies.
~ionization parameter Physical condition of ISM for SF galaxies at z=2-3 Lya emitters and Lyman break galaxies at z=2-3 Ha emitters at z=2.2-2.5 Shimakawa+15 (arxiv:1411.1408) Nakajima & Ouchi 14 ~oxygen abundance Higher ionization parameter of LAEs and LBGs at a given metallicity Positive correlation between electron density (n e ) and specific SFR These results indicate that physical condition of ISM in z=1-2 galaxies can be different from that of the local galaxies
Offset of the intensity ratios of nebular emission lines Well-known nebular emission lines in HII regions: Ha, Hb, [OII], [OIII], [NII], [SII] and so on BPT diagram: This diagram is frequently used to identify ionizing source Steidel+14 Newman+14 AGN Red: core, blue: edge, green: between them Star-forming The offset is NOT necessarily due to AGN. The line intensity ratios of z~2 galaxies show the offset from those of local galaxies.
Impact on metallicity measurement and SF/AGN classification The ratio of [NII]/Ha is frequently used to estimate oxygen abundance N2 -> O/H Underestimated? BPT diagram Dependence on q, n e, and radiation field AGN Dependence on q Maier+14 Star-forming If ionization parameter (q) of high-z galaxies is higher than local galaxies, oxygen abundance can be underestimated. Kewley+13 The diagnositcs with nebular emissions may not be applicable to high-z galaxies
Is the redshift evolution of M-Z relation real? Lower metallicity for galaxies with a given stellar mass at higher redshifts Zahid+14 With the evidences that the ISM condition may evolve with redshift, are the current measurements of metallcity reliable?
Our approach First of all, we would like to better understand the HII regions of star-forming galaxies at z>1 where useful nebular emission lines to study galaxy properties are observed. - Basic properties of the ISM in high-z galaxies: electron density, (electron temperature), ionizing source, ionization state, dust attenuation, and so on - How different in the line intensity ratios, compared with local galaxies
FMOS spectroscopy for [OII] emitters at z~1.5 in the SDF Subaru/FMOS spectroscopy (S12A-028 and S14A-018) [OII] emitters at z=1.47 (NB921) and z=1.62 (NB973) selected by S-Cam (Ly+07,12) HR(R=2200) in J-long Hb and [OIII]ll4959,5007 HR in H-long Ha, [NII]l6584, and [SII]ll6716,6731 Stellar mass from the SED fitting to the NUV-to-NIR photometry (Ly+12) Filled: confirmed galaxies NB921 [OII] NB973 [OII] 118 galaxies are spectroscopically confirmed (115 Ha / 45 [OIII] / 13 Hb / 14 [NII] / 5 [SII]) [OII] luminosity known for all of the galaxies the brightest galaxy ~3 orders of mag.
Composite spectra of [OII] emitters at z~1.5 Luminosities of 6 lines ([OII], Hb, [OIII], Ha, [NII], and [SII]) are available
What information do the nebular emission lines provide us with? Steidel+14 Ha (l6563) and Hb (l4861) Recombination lines (Balmer series) from hydrogen: directory reflect the number of ionizing photon - Luminosity: star formation activity - Intensity ratio of Ha/Hb: dust attenuation [OII](l3727), [OIII] (ll4959,5007), [NII] (l6584), and [SII] (ll6716,6731) Forbidden lines from ionized metal in HII regions: excited by collision with electron and then de-excited by radiation depends on electron temperature, electron density, and ionization state - [OII] luminosity: star formation activity (ionization potential ~ 13.6 ev) - Intensity ratio of [OII] or [SII] doublet: electron density - Intensity ratio of [OIII]/[OII]: ionization state - Intensity ratio of R23, ([OII]+[OIII])/Hb: oxygen abundance (O/H), ionization parameter
Intensity ratio Electron density Contours: local SF galaxies (from SDSS) Black filled circles: individual galaxies at z~1.5 [SII]6716/6731 Stack for Sample-1 Stack for Sample-2 Stack for Sample-3 10 2 10 3 Electron density [cm -3 ] An electron density (n e ) of z~1.5 star-forming galaxies is consistent with typical electron densities for local galaxies, i.e., n e ~10-100 cm -3.
Balmer decrement (the ratio of Ha to Hb) method of dust correction Yes Hα/Hβ available? No Hα/Hβ>2.86? Use the relation A(Hα) and M stellar (Garn&Best 2010) Yes estimate A(Hα) from Balmer decrement No A(Hα)=0 (no dust) The dust extinction curve of Calzetti+2000 is assumed. Under the assumption of Te=10 4 K and n e =100cm -3 for Case B recombination, the intrinsic Ha/Hb is 2.86 (Osterbrock 1989)
BPT diagram Offset of the line ratios, which is consistent with other studies
BPT diagram Shift toward large [OIII]/Hb ratio for z~1.5 galaxies Offset of the line ratios, which is consistent with other studies
High ionization parameter of [OII] emitters at z~1.5 [OIII]/[OII]: indicator of ionization parameter R23=([OII]+[OIII])/Hb (i.e., oxygen abundance) Contours: local SF galaxies, AGN, and composite objects (from SDSS) Nakajima & Ouchi 14 Large [OIII]/[OII] ratios, compared to local SF galaxies Comparable to LAEs/LBGs at z=2-3 [OII] emitters at z~1.5 have large [OIII]/[OII] ratio of >~1, suggesting they have large q
Dependence of [OIII]/[OII] on stellar mass Contours: local SF galaxies (from SDSS) Less massive galaxies have larger [OIII]/[OII] ratios May be due to higher ssfr for less massive galaxies
What causes the high ionization parameter?
Three possible mechanisms Hydrogen density Shape of ionizing radiation field Tasca+14 Geometry of nebular gas Nakajima & Ouchi 14
Three possible mechanisms Hydrogen density Shape of ionizing radiation field Tasca+14 Geometry of nebular gas Nakajima & Ouchi 14 Harder radiation field by young stellar populations The increasing number of ionizing photon from massive stars
Main sequence of star-forming galaxies [OII] emitters at z~1.5 follow the main sequence of star-forming galaxies x 20-50
How are the local star-forming galaxies with high SFR? Only quite a small number of local star-forming galaxies have SFRs comparable to [OII] emitters at z~1.5
SDSS galaxies with high SFR Difference in metallicity Not necessarily have large line ratios of [OIII]/[OII] Harder radiation fields at high-z?
Summary Hayashi et al., 2015, PASJ (Subaru Special Issue), submitted Subaru/FMOS Near-IR spectroscopy for 118 [OII] emission line galaxies at z~1.5 in the Subaru Deep Field Six nebular emission lines in rest-optical: [OII], Hb, [OIII], Ha, [NII], and [SII] Use line ratios to investigate physical conditions of the ISM Strong [OIII] emission line - [OIII]/[OII] ratios larger than the local star-forming galaxies - Less massive galaxies have larger [OIII]/[OII] ratios (mass dependence) High ionization parameter - Harder radiation field by young stellar populations - The increasing number of ionizing photon from massive stars Maybe relatively higher N/O abundance ratio, compared to local galaxies A tight correlation between Ha and [OII] - [OII] can be a good indicator of star formation rate for galaxies at z~1.5 It is important to take account of ionization parameter and nitrogen abundance ratio, when metallicity is derived for high-z galaxies.